Bring Your Own Key Model
Our plugin operates on a Bring Your Own Key basis, which means it serves as a bridge between your WordPress site and Azure OpenAI services..
In order to use our plugin, you need to have an Azure account and purchase API credits from Azure.
With this approach, users are required to provide their own Azure API key to access the plugin's functionality. This ensures that users maintain full control over their API keys and guarantees access to the latest, most advanced language processing technology offered by Azure OpenAI.
How to Generate an Azure OpenAI API Key
Azure OpenAI Service is a powerful tool integrated into the Azure ecosystem. Here's how you can get started with it:
Azure OpenAI Service, despite being generally available, still requires users to register. Microsoft might restrict access based on availability and other eligibility criteria.
- Visit Azure OpenAI Service.
- Click on the
- Complete the form that appears. Make sure to use your work email address; personal emails like Gmail or Hotmail might decrease your chances of approval.
After submission, Microsoft will review your request. Once approved, they will inform you about the access.
Accessing Azure Portal
- Navigate to Azure Portal.
- Click on "Create a Resource".
- In the search bar, type "OpenAI". Select Azure OpenAI from the search results.
- Click on
Create Azure OpenAI. This action will initiate the creation of a resource for you.
- After the resource is created, select it. On the left menu, look for "Keys and Endpoint" under the Resource Management section.
- Here, you can find your Azure OpenAI API keys and endpoint. Click the copy icon to copy the API key.
Setting Up AI Power Plugin
- Go to the AI Power plugin on your WordPress dashboard.
- Navigate to
- From the dropdown list, select
Azure. This action will present four new fields:
- Azure API key
- Azure Endpoint
- Azure Deployment
- Azure Embeddings
The first three fields are mandatory, while the last one is optional (but required if you're using embeddings).
- Fill in your API key and endpoint.
- Now, navigate to Azure AI Studio and log in.
- From the left menu, click on
Models. This will display all the models available to deploy.
Not all models can be deployed; only those marked "Yes" under "Deployable" can be.
Deploying a Model
- From the models list, select a deployable model.
- Click the
Deploybutton at the top.
- In the modal window, specify the base model, model version, and deployment name.
- After deployment, copy the deployment name and paste it into the plugin under
For embeddings, select the
text-embedding-ada-002 model and deploy it. Once deployed, copy its name and paste it in the plugin's relevant field.
Once all fields are populated, click the
Save button in the plugin.
Before setting Azure as your provider in the plugin, you should be aware of its limitations:
- Unsupported Modules: Fine-tuning and Audio Converter.
Once you enable Azure:
- Speech to Text: This feature will be disabled because Whisper is not yet available on Azure. However, we plan to integrate it once it becomes available.
- Moderation: Azure offers its own content filtering, so our plugin will automatically disable the moderation feature.
- Image Generator: Azure only allows up to 5 images maximum whereas OpenAI allows up to 10 images.
Frequently Asked Questions (FAQs)
1. Why is GPT-4 not in the Azure model list?
Answer: You need to fill out another form to get on the GPT-4 waitlist. Apply for GPT-4 waitlist here.
2. What is the price for using Azure OpenAI Service?
Answer: Pricing is based on a pay-as-you-go consumption model, with a price per unit for each model. For detailed pricing, check Azure OpenAI Service pricing.
3. What criteria does Azure use to evaluate and approve submitted use cases?
Answer: Azure evaluates submitted use cases based on several factors, including:
- Responsible use of the service—making sure the models are used as intended.
- Capacity—ensuring they can onboard and serve customers adequately.
4. How is Azure OpenAI Service different from the OpenAI API provided by OpenAI?
Answer: Azure OpenAI Service combines the capabilities of OpenAI API with Azure's enterprise-level security, compliance, and regional availability.
5. What models are available through Azure OpenAI Service?
Answer: Azure OpenAI Service offers access to GPT-3, Codex, and Embeddings models. Dall·E 2 is available only to invited customers. Learn more about the models here.
6. What is the service-level agreement (SLA) for Azure OpenAI Service?
Answer: Azure guarantees that Azure OpenAI Service will have an uptime of at least 99.9%. For more detailed information, view the SLA details.